DESCRIPTION: Magnetic Resonance Imaging is a technology that has good spatial resolution, avoids ionizing radiation, and can extend to a comprehensive examination. It also has improved dramatically in the quality of high speed acquisitions providing high resolution data. A major difficulty in bringing the new acquisition methods to full fruition in clinical use is a practical one. The high speed fine resolution data collections provide such a flow of data that it is impractical to analyze all the data for routine clinical applications. This study aims to perform a detailed evaluation on an extensive series of training data to characterize data reduction, information extraction, and error propagation of data reduction techniques.